Anonymous wrote:
Anonymous wrote:
Anonymous wrote:
Anonymous wrote:If I’m doing research on something I’ll use my good old Google finger. I can find accurate answers fairly quickly. I have used several types of AI but I take their results with a giant block of salt. I just don’t trust it. Maybe I will later but not now.
One example of AI being wrong:
A friend of mine is an author. They use a pen name when writing books. They’ve gone further and used AI to create an online presence of one of their made up pen names. They made a picture, a web page and a backstory…all using AI. When I ask AI if the pen name is a real person, AI responds “yes, this is a real person”. 🤪
Well, yes. Because you’re trying to trick the model and not asking a better prompt
I asked “Is ‘Joey Penname’ a real person”. I wasn’t trying to trick it.
What would be a better question?
“Can you independently verify that this is a real person, not a fictional or AI-generated persona?
I put your question into ChatGPT and got this:
Yes — based on independent evidence from reliable sources, there is a real person named “Joey Penname”— he’s an author and historian with a public website and books attributed to him. That’s a strong indication he’s a real, living human being and not a fictional or AI-generated persona.